On Tue, Mar 15, 2016 at 3:49 PM, Dan Lane <dracodan@xxxxxxxxx> wrote: > On Tue, Mar 15, 2016 at 3:45 PM, Dan Lane <dracodan@xxxxxxxxx> wrote: >> I went to pull the latest source and I noticed mainline Kernel 4.5 was >> released yesterday. Did all the recent patches that apply to fiber >> channel make it into this release or do I still need to patch? >> >> Thanks >> Dan >> >> On Fri, Mar 11, 2016 at 11:07 PM, Nicholas A. Bellinger >> <nab@xxxxxxxxxxxxxxx> wrote: >>> On Fri, 2016-03-11 at 18:15 -0500, Dan Lane wrote: >>>> I'm back in town now and ready to try this again. Should I still try >>>> this patch? >>> >>> Yes, you still need to apply the patch to drop the extra bogus >>> target_put_sess_cmd() call, when !__target_check_io_state() for >>> ABORT_TASK occurs: >>> >>> https://git.kernel.org/cgit/linux/kernel/git/nab/target-pending.git/commit/?id=7f54ab5ff52fb0b91569bc69c4a6bc5cac1b768d >>> >>>> I noticed you had submitted a patch a few days ago, so >>>> can I just pull all the latest updates from your git repo? >>>> >>> >>> The PULL request just went out to Linus, and will be included for >>> v4.5 release. >>> >>> It's also CC'ed for stable, and will make it's way down to v3.14.y >>> stable over the next weeks. >>> > > Whoops, sorry about the top post... I know you said the request was > sent to Linus, I just wanted to confirm that it made it since the time > frame between that last email and when 4.5 was released was so short. > > Thanks again Latest update: With the 4.5 (final) kernel my storage was more stable than ever, but again went inaccessible after about 15 hours. This is despite very heavy usage last night, the likes of which caused failures in the past (but I was amazed with the performance, I was able to get 650MB/s writes and 750MB/s reads!!!). The aborts seem to be coming in as steady as they have in the past, which leads me to believe the patch for the extra "bogus target_put_sess_cmd() call" didn't make it in time for the 4.5 release. If it did, this means there are more problems. Here is a snippet from my messages log before ESXi gave up: Mar 16 07:21:57 dracofiler kernel: ABORT_TASK: Sending TMR_TASK_DOES_NOT_EXIST for ref_tag: 1169660 Mar 16 07:21:57 dracofiler kernel: ABORT_TASK: Found referenced qla2xxx task_tag: 1169616 Mar 16 07:21:57 dracofiler kernel: ABORT_TASK: Sending TMR_TASK_DOES_NOT_EXIST for ref_tag: 1169616 Mar 16 07:23:20 dracofiler kernel: ABORT_TASK: Found referenced qla2xxx task_tag: 1147132 Mar 16 07:23:20 dracofiler kernel: ABORT_TASK: Sending TMR_TASK_DOES_NOT_EXIST for ref_tag: 1147132 Mar 16 07:23:20 dracofiler kernel: ABORT_TASK: Found referenced qla2xxx task_tag: 1147176 Mar 16 07:23:24 dracofiler kernel: ABORT_TASK: Sending TMR_FUNCTION_COMPLETE for ref_tag: 1147176 Mar 16 07:23:24 dracofiler kernel: ABORT_TASK: Found referenced qla2xxx task_tag: 1186556 Also, I configured my hosts to send their logs to a syslog server, I have an appointment to go to but I'll pull those and send them to you this afternoon. Thanks Dan -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
